Re: woody: Remaining RC bugs in base system

[Sam Hartman <hartmans@debian.org>]

>>>>> "Mark" == Mark Brown <broonie@sirena.org.uk> writes:

    Mark> On Thu, Jan 24, 2002 at 09:41:01PM -0500, Matt Zimmerman
    Mark> wrote:
    >> Such is the way of the freeze, I think.  This will have to
    >> happen at some point when new versions of base packages stop
    >> flowing into testing.  Either unstable will have to be frozen
    >> as well (in which case, we have missed part of the point of
    >> testing) or packages will have to be compiled against woody.

    Mark> If packages that are frozen just stay frozen and don't get
    Mark> people starting to upload incompatible new versions this
    Mark> wouldn't be a problem.  If packages are going to start
    Mark> getting new incompatible versions then aren't we going to
    Mark> start getting back to an old-style freeze but without the
    Mark> ability to upload to frozen (making bug fixing a bit hard)?

There have been several past discussions of this here, including mail
from aj.  The consensus of those discussions is that libraries MUST
NOT change in incompatible ways (soname change, ABI change) once
frozen.  Aj actually made a stronger statement and said that people
should not upload base or standard libraries at all once they were
frozen (presumably unless they were also sending the library to
woody-proposed-updates or semthing).

So a new ncurses should not have been uploaded, especially if it is
incompatible, unless it is targeted for woody.